Studies on correlated responses of some yield components in two rice hybrids

Due to genetic correlations, selection on one trait might lead the plant breeders unexpected responses in non-selected characters. Or this can be used as a tool for further breeding activities. To study this phenomena, two rice crosses have been taken as material (Fero/Zenchuzry; Solnecnyt/Iz160). In each F2 population 10 plants have been selected for the following characters: the most productive; the shortest plant (their yield were over average); the longest panicle length; the highest tillering. Beside this, a randomly selected sub-population from each population have been selected and all have been transplanted next year as F3 generation. After routine observations or measurements (grain yield, number of days to flowering; culm length; panicle length; numbers of tillers; thousand kernel weight; total number of panicle; number of sterile spikelet and number of fertile spikelet per panicle), classical quantitative genetically parameters have been calculated. Thereafter expected and realized direct genetically responses and expected and realized correlated responses have been estimated. Expected and realized correlated responses were in the same (positive) direction for following selections in both hybrids: "Grain Yield - number of tiller", "Plant Height - thousand-kernel weight and panicle length", "Panicle Length - number of fertile spikelet per panicle", "Number of Tiller - grain yield". Those results are proving that, a successful selection of one economically important trait might cause a disappointment to the breeder due to unexpected correlated response in some traits. Therefore estimation of expected correlated responses might be quite useful to the breeders. (Résumé d'auteur